Two subclasses of lung squamous cell carcinoma with different gene expression profiles and prognosis identified by hierarchical clustering and non-negative matrix factorization

Abstract

Current clinical and histopathological criteria used to define lung squamous cell carcinomas (SCCs) are insufficient to predict clinical outcome. To make a clinically useful classification by gene expression profiling, we used a 40 386 element cDNA microarray to analyse 48 SCC, nine adenocarcinoma, and 30 normal lung samples. Initial analysis by hierarchical clustering (HC) allowed division of SCCs into two distinct subclasses. An additional independent round of HC induced a similar partition and consensus clustering with the non-negative matrix factorization approach indicated the robustness of this classification. Kaplan-Meier analysis with the log-rank test pointed to a nonsignificant difference in survival (P = 0.071), but the likelihood of survival to 6 years was significantly different between the two groups (40.5 vs 81.8%, P = 0.014, Z-test). Biological process categories characteristic for each subclass were identified statistically and upregulation of cell-proliferation-related genes was evident in the subclass with poor prognosis. In the subclass with better survival, genes involved in differentiated intracellular functions, such as the MAPKKK cascade, ceramide metabolism, or regulation of transcription, were upregulated. This work represents an important step toward the identification of clinically useful classification for lung SCC.